DNS failover: switch to a backup target automatically

With DNS failover, IPv64 health checks monitor your service. If it fails, IPv64.net automatically switches the A and AAAA records of your domain to a backup target – another IP, another record or an offline page – and, if you like, back again as soon as the service is running.

How DNS failover works

A policy connects two things:

• Monitors – one or more of your health checks. Their state decides whether the policy triggers.
• Rules – which record should point to which backup target on failure.

When the policy triggers (failover), IPv64.net takes the affected original records out of DNS and creates replacement records with the target from the rule instead. On fallback the replacement records are deleted and the original records become active again. A and AAAA records are supported, i.e. IPv4 and IPv6. DNS failover is also available in the free account.

Health check(s) ──► Policy ──► Rule: www.yourdomain.ipv64.net  A  192.0.2.10
   Down               │                      │ Failover
                      │                      ▼
                      └───────► Backup: www.yourdomain.ipv64.net  A  192.0.2.20  (TTL 10 s)

Setting up a policy

Prerequisite: you have at least one health check monitoring your service (see Monitors), and a domain with an A or AAAA record (see DynDNS domains).

1. Create a policy
   Open ipv64.net/dns_failover and create a new policy.
2. General settings
   Under “General Settings” enter a name and choose the mode under “Automated DNS Failover” (see Failover modes).
3. Add monitors
   Under “Monitor Healthcheck | Failover Trigger” select a health check, for the first monitor also AND or OR, optionally “Negate”, and click “Add Healthcheck to Policy”.
4. Add rules
   Under “DNS Failover Rules” select the record and the backup target.
5. Choose notifications
   Under “Notification integration” define who is informed on failover and fallback.

As long as a monitor or rule is missing, the policy shows “Pending”. As soon as there is at least one monitor and one rule, it becomes “Active” after a few seconds.

Limits per account class

Current values and all other limits: Account & limits.

Monitors: health checks as triggers

Each monitor is a health check from your account. You can choose the checks of the health checks – online, local and hybrid. Until Healthchecks 1.0 is shut down on 22 November 2026, the checks of the previous health check page work too; both share the same token, so a check you move over stays a monitor. Only two states count for failover: Up (OK) and Down. Paused checks and checks that are partially down, flapping or dependent count neither as failure nor as recovery; a deleted health check does not count either.

If a check is a monitor of a policy, its detail page in the health checks says so. Deleting it shows a warning with the policy name – then enter another health check as monitor there.

AND and OR linking

You choose the linking with the first monitor; it applies to all monitors of the policy.

Negate

“Negate” inverts a monitor: it counts as failed when the health check is Up, and as healthy when it is Down. This lets you switch deliberately, e.g. with a maintenance check. In the monitor list an inverted monitor is marked “NAND” or “NOR”.

Rules and backup targets

A rule names a record (A or AAAA) from your domains and its backup target. Each record can only be in one policy. There are three kinds of targets:

Fixed IP address

You enter an IP. It must match the record’s address family: IPv4 for A records, IPv6 for AAAA records.

Target via record key

Instead of a fixed IP, choose another record from your account under “Select predefined Dst.”. Its record key is stored; the IP is only read at the moment of failover. If the IP of this target record changes via a DynDNS update while failover is active, IPv64.net updates the replacement record automatically. Ideal when your backup server has a dynamic IP itself.

If the target record is an AAAA record that only contains the host part (starts with ::), it is combined with the stored IPv6 prefix of its domain.

Offline page

You can also choose an offline page as the target – the “Default IPv64.net Offline Page” or your own. Visitors then see a notice page while your service is down. Details: Offline pages.

Failover modes and fallback rule

Under “Automated DNS Failover” you choose what happens automatically:

Fallback rule

The option “All Monitoring-Rules must be healthy for Fallback” defines when to switch back automatically:

• checked: only when all monitors are Up again (with Negate: Down),
• unchecked: as soon as at least one monitor is healthy again – with AND as well as OR.

For a stable way back after an OR failover the checkbox is usually worth it: otherwise the policy may switch back while another monitor is still Down.

Manual failover, fallback and pause

• Manual failover: With an active policy (“Active”) you can trigger the failover by hand. The policy briefly shows “Manual Failover”, then “Manual Failed Over”.
• Manual fallback: When the policy has switched (“Failed Over”), you can switch back by hand – in every mode. Status briefly “Manual Fallback”.
• Pause: An active or pending policy can be paused (“Paused”). It then checks nothing and does not trigger. When resumed, it is active again after a few seconds. Pausing is not possible during a failover.

With “Only manual Failover” you use DNS failover as a tool: the health checks notify you, you decide yourself and switch with one click.

Notifications

On every failover (“DNS Failover Policy … TRIGGERED”) and every fallback (“… FALLBACK”), IPv64.net sends a message to the selected integrations. Under “Notification integration” you can select several; the default is your account’s e-mail address (“Account E-Mail (default)”). “Disable all Notifications” turns messages off for this policy. Both events are also written to the account log.

DNS failover and CDN64

DNS failover and CDN64 cannot be combined for the same name. When a name is switched on in CDN64, it points to the edge servers in DNS and its own A/AAAA records are dormant. DNS failover therefore does not offer these records for selection. Conversely, a name with an active DNS failover cannot be switched on in CDN64 (message DNS failover is active for this name.).

CDN64 has its own failover: four edge servers, of which a failed one is removed from DNS automatically, and in modes 1/2 optionally several Node64 instances as backup (see load balancing and failover).

Process, interval and TTL

• Check interval: The failover service checks all policies at short intervals, roughly every 20 seconds.
• TTL of replacement records: 10 seconds so that resolvers ask again quickly. On fallback the TTL of your account class applies again.
• Caches: Resolvers that already cached the original record keep serving it until its TTL expires. Short TTLs (see account classes) make failover faster.
• Deleted original records: If a rule’s record is missing, failover skips that rule.

You can also manage policies, monitors and rules via API v2 (section “DNS-Failover”); see also API.

Found a mistake or missing something? Tell us in the v64.tech forum. · Updated: 2026-11-04